THE Brassica genus contains the largest number of cultivated. The Fate of Chromosomes and Alleles in an Allohexaploid Brassica Population

Transcription

1 INVESTIGATION HIGHLIGHTED ARTICLE The Fate of Chromosomes and Alleles in an Allohexaploid Brassica Population Annaliese S. Mason,*,1 Matthew N. Nelson,, Junko Takahira, Wallace A. Cowling, Gustavo Moreira Alves,*, Arkaprava Chaudhuri,*, ** Ning Chen,* Mohana E. Ragu,* Jessica Dalton-Morgan,* Olivier Coriton, Virginie Huteau, Frédérique Eber, Anne-Marie Chèvre, and Jacqueline Batley*,1 *School of Agriculture and Food Sciences and Centre for Integrative Legume Research, The University of Queensland, Brisbane, Queensland 4072, Australia, School of Plant Biology and The UWA Institute of Agriculture, The University of Western Australia, Crawley 6009, Perth, Western Australia Australia, Universidade Federal de Minas Gerais, 6627, Pampulha Belo Horizonte MG, , Brazil, **Department of Biotechnology, National Institute of Technology, Durgapur, West Bengal , India, and Unité Mixte de Recherche Institut de Génétique, Environnement et Protection des Plantes, Institut National de la Recherche Agronomique, BP35327, Le Rheu Cedex, France ABSTRACT Production of allohexaploid Brassica (2n = AABBCC) is a promising goal for plant breeders due to the potential for hybrid heterosis and useful allelic contributions from all three of the Brassica genomes present in the cultivated diploid (2n = AA, 2n = BB, 2n = CC) and allotetraploid (2n = AABB, 2n = AACC, and 2n = BBCC) crop species (canola, cabbages, mustards). We used highthroughput SNP molecular marker assays, flow cytometry, and fluorescent in situ hybridization (FISH) to characterize a population of putative allohexaploids derived from self-pollination of a hybrid from the novel cross (B. napus 3 B. carinata) 3 B. juncea to investigate whether fertile, stable allohexaploid Brassica can be produced. Allelic segregation in the A and C genomes generally followed Mendelian expectations for an F 2 population, with minimal nonhomologous chromosome pairing. However, we detected no strong selection for complete 2n = AABBCC chromosome complements, with weak correlations between DNA content and fertility (r 2 = 0.11) and no correlation between missing chromosomes or chromosome segments and fertility. Investigation of next-generation progeny resulting from one highly fertile F 2 plant using FISH revealed general maintenance of high chromosome numbers but severe distortions in karyotype, as evidenced by recombinant chromosomes and putative loss/duplication of A- and C-genome chromosome pairs. Our results show promise for the development of meiotically stable allohexaploid lines, but highlight the necessity of selection for 2n = AABBCC karyotypes. THE Brassica genus contains the largest number of cultivated crop species of any plant genus (Dixon 2007). Six major crop species are B. rapa (Chinese cabbage, turnip), B. oleracea (cabbage, cauliflower, broccoli), B. nigra (black mustard), B. napus (canola, rapeseed), B. juncea (Indian mustard, leaf mustard), and B. carinata (Ethiopian mustard). These six species share a unique genomic relationship: progenitor diploid species B. rapa (2n =AA),B. oleracea (2n =CC), Copyright 2014 by the Genetics Society of America doi: /genetics Manuscript received November 10, 2013; accepted for publication February 13, 2014; published Early Online February 20, Supporting information is available online at doi: /genetics /-/dc1. 1 Corresponding authors: John Hines Bldg. 62, Centre for Integrative Legume Research, School of Agriculture and Food Sciences, The University of Queensland, Brisbane 4072, Australia. annaliese.mason@uq.edu.au; j.batley@uq.edu.au and B. nigra (2n = BB) gave rise to the allotetraploid species B. juncea (2n = AABB), B. napus (2n =AACC),andB. carinata (2n = BBCC) through pairwise crosses (Morinaga 1934; U 1935). However, despite the fact that each pair of genomes coexists in an allotetraploid species, no naturally occurring allohexaploid species (2n = AABBCC)exists.In general,interspecific hybridization and polyploidy in plants are potent evolutionary mechanisms, allowing formation of new species with adaptation to a wider range of climatic conditions and greater hybrid vigor than their progenitor species (Otto and Whitton 2000; Leitch and Leitch 2008). Hence, production of an artificial allohexaploid in the agriculturally important Brassica genus could potentially give rise to new crop types with greater intersubgenomic heterosis (Zou et al. 2010) and tolerance of a wider range of environmental conditions than preexisting Brassica crops (Chen et al. 2011). Genetics, Vol. 197, May

2 Interest in producing an allohexaploid Brassica dates back to the 1960s when induction of somatic chromosome doubling of triploid ABC interspecific hybrids to form AABBCC allohexaploids was first carried out (Iwasa 1964). However, mixed results have been reported for success of these synthetic allohexaploid Brassica. Howard (1942) reported relatively high fertility over several generations of self-pollination of 2n = AABBCC plants derived from the cross B. rapa 3 B. carinata. However, Iwasa (1964) found poor meiotic stability in the same species cross that failed to improve even up to the F 5 selfing generation, resulting in loss of fertility and hence usefulness of this crop. Allohexaploid Brassica has also been created as a bridge to transfer disease resistance between B. nigra and B. napus (Sjödin and Glimelius 1989; Pradhan et al. 2010) to produce yellow-seeded B. napus by crosses between B. rapa and B. carinata (Meng et al. 1998; Rahman 2001) to transfer cytoplasmic male sterility from B. oleracea to B. juncea (Arumugam et al. 1996) and to resynthesize B. napus from B. rapa and B. carinata (Li et al. 2004). However, although these studies demonstrate the ease of production of allohexaploid Brassica from crosses between the diploid and allotetraploid species, meiotic stability and fertility was not assessed in these allohexaploid types. Somewhat increased genomic stability in allohexaploids derived from genotypes of B. rapa 3 B. carinata crosses has recently been demonstrated after several successive generations of selfing and selection for 2n = 54 chromosome complements (Tian et al. 2010). Mason et al. (2012) mimicked a natural evolutionary pathway for polyploid formation in Brassica by using unreduced gametes to produce an allohexaploid plant. B. napus and B. carinata were crossed in the first generation to create a hybrid with karyotype CCAB (Figure 1). An unreduced gamete from this hybrid combined with a reduced gamete from B. juncea (n = AB) to form a near-hexaploid (AABBCC) plant with 2n = 49 chromosomes, 5 short of the full allohexaploid complement (Mason et al. 2012) (Figure 1). Low-resolution molecular karyotyping using microsatellite markers indicated that the missing chromosomes were all monosomic deletions [i.e., one copy of each of five different homologous chromosomes was retained in the allohexaploid plant (Mason et al. 2012)]. This plant was highly fertile and provided plentiful selfpollinated progeny for the present study. We hypothesized that some lines produced from this allohexaploid may show 2n = 54, as well as demonstrating greater meiotic stability and fertility than lines produced through crosses between diploid and allotetraploid species followed by colchicine treatment to induce chromosome doubling. Materials and Methods Generation and growth of plant material and experimental populations Methods used in crossing B. napus 3 B. carinata to produce the first generation hybrid plant (2n = CCAB) are detailed in Mason et al. (2011). Crossing to generate the allohexaploid hybrid plant (hereafter referred to as A1 ; previously N1C1.J1-1) and preliminary cytological and molecular marker characterization of this plant are detailed in Mason et al. (2012) and further details are provided in Figure 1. The source of cytoplasm in A1 was B. napus, with B. carinata and B. juncea used as paternal parents in the formation of A1. The A genomes in B. juncea and B. napus are hereafter referred to as subgenomes A j and A n, respectively; the B genomes in B. juncea and B. carinata are hereafter referred to as subgenomes B j and B c, respectively; and the C genomes in B. napus and B. carinata are hereafter referred to as subgenomes C n and C c, respectively, following the naming conventions of Li et al. (2004). Seeds from self-pollination of hybrid plant A1 were sown and are subsequently referred to as the SP population. Data collected from each experimental plant in the SP population are summarized in supporting information, Table S1. We characterized subsets of plants for total DNA content (flow cytometry), for A- and C-genome transmission (Illumina Infinium Brassica 60K SNP array), and for fertility (sporad observations, pollen viability, and self-pollinated seed set). A set of second-generation plants resulting from one of the most fertile SP progeny was then characterized by chromosome counts, and three plants with.50 chromosomes were tested for genome inheritance using fluorescent in situ hybridization (FISH). The SP population was split into two groups. The first group (lines SP_001 to SP_050) were germinated in potting mix and grown to maturity in a controlled environmental room (16-hr photoperiod with a light intensity of 200 mmol m 22 s 21 at 18 /10 day/night) at The University of Western Australia (UWA), Perth, Australia. Lines SP_051 to SP_115 were germinated in potting mix and grown to maturity under air-conditioned glasshouse conditions at The University of Queensland (UQ), Brisbane, Australia. Self-pollinated seeds from plant SP_042 were germinated and grown under heated glasshouse conditions at the Institut National de la Recherche Agronomique (INRA), Le Rheu, France. DNA was extracted from leaf tissue of plants SP_001 to SP_050 and parent species controls usinganucleonphytopuregenomic DNA Extraction Kit (Illustra). DNA was extracted from leaf tissue of SP lines SP_051 to SP_115 using the Microprep method of Fulton et al. (1995). Flow cytometry Samples SP_001 to SP_050 and controls (parent species genotypes, CCAB hybrids, and plant A1) were assessed for DNA content using flow cytometry. Young leaf tissue (5 mg) from each plant was individually sampled in 1.2-ml eightstrip tubes containing a 3-mm tungsten carbide bead on ice, following the methods of Cousin et al. (2009). To each sample, 500 ml of cold lysis buffer (0.1 M citric acid and 0.5% v/v Triton X-100 stored at 4 ) was added. Samples were then placed at 220 for 10 min. Samples were shaken at 25 Hz for a total of 24 sec in a TissueLyser (Qiagen), with the rack orientation reversed at 12 sec to help evenly distribute the 274 A. S. Mason et al.

3 Figure 1 Generation of experimental Brassica populations. (1) Interspecific hybridization between B. napus and B. carinata in the first generation. (2) Hybridization between the B. napus 3 B. carinata hybrid and B. juncea in the second generation, with the B. napus 3 B. carinata hybrid producing an aneuploid unreduced gamete. (3) Selfpollination of the resulting plant A1 to give rise to the SP lines. (4) Self-pollination of plant SP_042 to give rise to second generation SP42 progeny. homogenization. Homogenate (120 ml) was transferred by multichannel pipette to 96-well filter plates [800-ml Unifilter microplate with a 25- to 30-mm pore size and long drip directors (Whatman)] sitting over collection plates, which were gently vacuumed to collect the filtrate. A reading plate with 2 ml RNase A (3 mg/ml) per well was set up, and 60 ml ofthefiltered samples was added to respective well locations. The reading plate was incubated at 37 for 30 min and treated with 150 ml propidium iodide (PI) staining solution (Roberts 2007) following the methods of Cousin et al. (2009). Nuclear DNA content was measured on a BD FACSCanto II (BD Biosciences) flow cytometer, using FACS Diva V6.1.1 operating software at The University of Western Australia s Centre for Microscopy, Characterization and Analysis. The PI dye was excited with a 488-nm (blue) laser and PI emission collected with a 585/42 ( nm) bandpass filter for 60 sec per sample or for 20,000 nuclei events (whichever came sooner). Samples that did not show a single clear peak in the first run (26 SP lines) were repeated up to five times. Lettuce (Lactuca sativa Grand Rapids ) was used as the internal standard for nuclear DNA content, and B. rapa, B. napus, andb. carinata were used as external standards. The L. sativa histogram peak was adjusted to fluorescence channel 200 at the beginning of each acquisition session. Experimental data were analyzed and coefficient of variation (CV) calculations performed using FlowJo V7.2.5 FC analysis software (Tree Star Inc.) to determine the CV, mean size of the gate and means for the standard and sample peaks (Cousin et al. 2009). Sample 2C DNA was calculated by (mean of sample peak/mean of standard peak)*2c DNA amount of the standard species (L. sativa: 5.5 pg) (Zonneveld et al. 2005). Ploidy x-value was calculated by 2C value/( ): (B. rapa AA-genome + B. carinata BBCC-genome, respectively). Pollen viability Flowers from all samples were collected, and anthers were dissected from flowers and placed on microscope slides with a few drops of 1% acetic acid carmine solution (1 g carmine powder in 45% glacial acetic acid solution). Pollen viability was assessed as described in Nelson et al. (2009): stained and swollen pollen grains (oval shape) were considered viable, and nonstained and shrivelled pollen grains were counted as nonviable. A minimum of 300 pollen grains were counted per plant. Sporad observations Buds were collected in 6:3:1 ethanol:chloroform:glacial acetic acid solution (Carnoy s II) in the morning and incubated for 24 hr at room temperature before being transferred to 50% ethanol at 4 for storage. Anthers were dissected out from buds on glass microscope slides and gently squeezed to release sporads into 1% acetic acid carmine solution. The number of sporads with one, two, three, four, five,orsixnucleipostmeiosis (monads, dyads, triads, tetrads, pentads, and hexads, respectively) was counted for each of 58 individuals in the SP population, with at least 300 sporads observed per plant. Chromosome counting Twenty-five plants resulting from self-pollination of SP_042 (second-generation self-pollinated plants) were assessed for Allohexaploid Brassica Genetics 275

4 chromosome number. Three to 10 (mode five) counts were performed for each plant on different chromosome spreads. Fresh root tips were collected into glass bottles containing 0.04% 8-hydroxyquinoline solution, incubated for 2 hr in the dark at room temperature, and then for another 2 hr at 4. Samples were fixed in cold ethanol:glacial acetic acid (3:1) solution for 48 hr at room temperature and then stored in 70% ethanol at 220 until used. Fixed root tips were rinsed with deionized water for min at room temperature to remove the fixative, then incubated in 0.01 M citrate solution (6 mm trisodium citrate dihydrate, 4 mm citric acid monohydrate, ph 4.5) for 15 min at room temperature, and then in 250 ml of enzyme solution [5% cellulase (Sigma-Aldrich) and 1% pectolyase (Sigma-Aldrich) in 5 ml citrate solution] for 30 min at 37. After incubation, the enzyme solution was removed with the aid of a Pasteur pipette, and roots were soaked in deionized water for 30 min. Washed meristem root tips were transferred to slides using a Pasteur pipette. One drop of DAPI (2 mg/ml of 49,6 diamidino-2-phenylindole in McIlvaines citric buffer, ph 7; blue emission)(vector Laboratories) was added. Fluorescence images were captured using a CoolSnap HQ camera (Photometrics) on an Axioplan 2 microscope (Zeiss) and analyzed using MetaVue (Universal Imaging). FISH FISH was carried out according to protocols detailed in Mason et al. (2010), at INRA, Le Rheu, France. The B genome was detected using Texas Red-labeled B. nigra DNA as a probe (GISH, Genomic In Situ Hybridisation), and the C genome was detected using Alexa 488-labeled BAC BoB014O06 to highlight a C-genome-specific repetitive sequence distributed over the nine C-genome chromosomes (Howell et al. 2002; Leflon et al. 2006). Three plants with chromosome counts of 2n = 54 were chosen for FISH. Molecular karyotyping using the Illumina Infinium Brassica 60K SNP chip A SNP chip with 52,157 SNPs designed for B. napus (Illumina Infinium 60K bead array) was used to assess allele inheritance in SP_051 to SP_115, as well as parental controls [see Mason et al. (2014) for details of SNP data analysis]. Chip hybridization was carried out according to the manufacturer s instructions at The University of Queensland, Brisbane, Australia, and chips were scanned using a HiScanSQ (Illumina). Data were visualized using the genotyping module of Genome Studio V (Illumina), and SNP locations were sorted with reference to the publically available sequenced B. rapa genome (X. Wang et al. 2011) and B. oleracea genome ( Statistical analyses and genetic segregation hypotheses R version (The R project for statistical computing) base and sm packagewereusedtogeneratehistogramsand density distributions. Pearson s x 2 test was used to assess whether allele inheritance followed Mendelian expectations for an F 2 selfing population. For each locus at which both an A j and an A n allele was present in the A1 allohexaploid plant (A j /A n ), a 1: 2:1 allelic segregation ratio was predicted for A n /A n :A j /A n :A j /A j in the progeny. Likewise, for each locus at which both a C n and a C c allele was present in the allohexaploid plant A1 (C n /C c ), a 1:2:1 allelic segregation ratio of C n /C n :C n /C c :C c /C c was predicted in the segregating F 2 population. For univalent chromosomes A j 2, A j 6, and A j 9, where no B. napus (A n ) homologous chromosome partner was present in the A1 hybrid, gametes from A1 would be expected to carry each chromosome at a frequency of 50%. As two parent gametes united in a self-pollination event to produce the SP progeny, 75% of plants should have at least one copy of the univalent B. juncea alleles at each locus (1:2:1 ratio for two copies of A j :one copy of A j :no copies of A j ). Hence, a 3:1 ratio of presence:absence was expected for A j alleles for chromosomes A j 2, A j 6, and A j 9. Results Flow cytometry Of the 44 plants assessed for DNA content using flow cytometry in the SP population (Table S1), 43 had DNA contents ranging from 82 to 97% of the predicted DNA content for a 2n = AABBCC allohexaploid (average 88%) and formed a normal distribution about the A1 plant DNA content (Figure 2). A single plant (SP_016) had a DNA content of 28% of the predicted allohexaploid DNA content, approximately the same as B. rapa (2n = AA). Molecular karyotyping using the Illumina Brassica 60K SNP chip for the A and C genomes A total of 7651 of the 52,157 A- and C-genome SNPs assessed were polymorphic between the B. juncea (A j )andb. napus (A n ) A genomes or B. napus (C n )andb. carinata (C c )C genomes in the parent cultivars: 2667 in the A genome (A j / A n SNPs) and 4984 in the C genome (C n /C c SNPs). The number of SNPs per chromosome ranged from 123 to 1283 (average 402 per chromosome). On the basis of the SNP molecular karyotyping, we determined if chromosomes from B. juncea (A j alleles), B. napus (A n and C n alleles), and B. carinata (C c alleles) were absent (all alleles from that subgenome absent for that chromosome), present (all alleles from that subgenome present for that chromosome), or partially present (some alleles from that subgenome absent and some present for that chromosome, such as would result from homologous recombination between A n and A j chromosomes or between C n and C c chromosomes). Absence of both parental alleles (i.e.,botha j and A n or both C n and C c ) was assumed to result from chromosome loss if a whole chromosome was missing and to result from nonhomologous recombination if only part of a chromosome was missing. B. napus chromosomes A n 2, A n 6, and A n 9 were missing from the allohexaploid hybrid A1 : these B. napus chromosomes were not transmitted to A1 from the C n C c A j B c parent and hence were also absent in the SP population. A previously 276 A. S. Mason et al.

5 Figure 2 Approximate DNA content determined by flow cytometry in 44 SP plants derived from a plant (A1) resulting from the cross (B. napus 3 B. carinata) 3 B. juncea. A DNA content of 1.0 represents the expected value for a 2n = 6x = AABBCC allohexaploid, produced by summing values from B. rapa (2n =2x = AA) and B. carinata (2n =4x = BBCC). characterized A n 7-C n 6 interstitial reciprocal translocation comprising about 1/3 of both chromosomes (Osborn et al. 2003) also existed within the B. napus cultivar but not in the B. juncea/b. carinata A j and C c genomes, as evidenced by haplotype block discontinuity and disrupted allelic inheritance in this region of the SP population. These events accounted for the vast majority of missing chromosome fragments in the 71 plants genotyped in the SP population (Figure 3): chromosomes A7 and C6 showed 16 and 12% loss of both parental alleles, respectively, across the population. The next largest percentage loss of both parental alleles (1 2%) was observed for chromosomes A1 and C2 (Figure 3). Part of chromosome A n 10 in the B. napus parent line exhibited two-locus amplificationpatternsacrossallsnps(asassessedbysnpmarker cluster plots) for a distance of 5 Mb at the distal end of chromosome A10. Two-locus patterns were not observed in the B. juncea A j 10 chromosome in the same region. This putative duplication in A n 10 could not be readily matched to a haplotype pattern present in any other part of the genome. Of the 71 SP progeny genotyped, 24 (32%) had complete complements of A- and C-genome alleles from at least one parent. For those chromosomes absent from the A1 hybrid (A n 2, A n 6, and A n 9), their homolog equivalents from the B. juncea parent (A j 2, A j 6, and A j 9) were present. In these cases, it was not possible to determine if one or two copies of each A j 2, A j 6, or A j 9 allele was present. Allele transmission and genomic bias in the A and C genomes Allelic inheritance for most chromosomes followed normal Mendelian expectations for segregation at heterozygous loci (Figure 3). For loci that were heterozygous in the nearhexaploid parent, a 1:2:1 ratio of codominant parental alleles was expected in the self-progeny. For loci present as a single copy in the allohexaploid hybrid A1 (i.e., part of univalent chromosomes), a dominant-type ratio of 3:1 present:absent was expected in the self-progeny. In total, 877 SNP loci were fixed to homozygosity for either B. napus or B. carinata parental alleles in the allohexaploid hybrid A1 (and subsequently in the SP population) due to homologous recombination between the C n and C c chromosomes during meiosis in the C n C c A n B c hybrid (Figure 1). Regions of homozygosity fixed by homologous recombination in the B. napus 3 B. carinata hybrid were present on every chromosome. Of the remaining 6774 loci, 626 (9.2%) showed distorted segregation toward retention of either parental allele at P, 0.05 (Pearson s x 2 test for count data), with 28 loci (0.4%) significant at P, All loci with significant distortion at P, were present in two clusters on chromosomes A7 and A10. The bottom part of B. napus A n 7 that represented the small telomeric region of the chromosome adjacent to the interstitial A7/C6 translocation region was retained more often than predicted by chance. B. napus alleles were also preferentially retained for the putative duplicated region on chromosome A n 10. In addition, 39% of the loci showing segregation distortion were present on B. juncea univalent chromosome A j 9. The univalent chromosome A j 9 from B. juncea was retained more often than predicted, although there was no such effect for the other univalent chromosomes A j 2 and A j 6 (Figure 3). Taking into account only loci segregating for C n and C c alleles, C-genome allele inheritance included a slight bias toward inheritance of a section of B. napus C n 2, retention of B. carinata alleles in a section of C c 4, bias toward presence of either B. napus or B. carinata C7 alleles for different parts of the chromosome and bias toward presence of B. carinata alleles in part of C8. Sections of chromosomes A3, A4, and A5 showed bias toward retention of B. juncea alleles, and the top of chromosome A8 and the bottom of A10 showed bias toward retention of B. napus alleles (Pearson s x 2 test, P, 0.05). No genome-wide bias toward inheritance of either alleles from a particular subgenome or alleles from a particular species was observed. Allohexaploid Brassica Genetics 277

6 Figure 3 Inheritance of A-genome alleles (A j and A n from B. juncea and B. napus, respectively) and C-genome alleles (C c and C n from B. carinata and B. napus, respectively) in 71 SP progeny of a hybrid plant derived from the cross (B. napus 3 B. carinata) 3 B. juncea. Fertility of the SP generation Pollen viability was assessed for 109 SP individuals. There was no significant difference in pollen viability between the two groups of the SP population grown at UWA and UQ (Student s t-test, P. 0.05). Pollen viability ranged from 0 to 92% across the SP population, with an average of 39% (Figure 4). A bimodal distribution for pollen viability was weakly suggested by the density distribution (Figure 4). Of the 109, 12 individuals were male-sterile (11%). The number of seeds per plant harvested from 109 progeny in the SP population ranged from 0 to 583, averaging 68 and with a median of 28 (Figure 5). Thirteen plants (12%) set no seeds, 34 plants (31%) set,10 seeds, and 28 plants (26%) set.100 seeds (Figure 5). There were no obvious relationships between genomic structure, including missing chromosomes, pollen viability, or self-pollinated seed set in the SP population. Sporad observations Fifty-eight individuals in the SP population were assessed for sporad production and produced an average of 95.2% tetrads, 1.3% triads, 1.9% dyads, and 1.5% monads. Tetrad production ranged from 61.3 to 99.7%, with a median of 96.7%. One (sterile, no self-seed produced) plant produced 36.2% dyads. The next highest dyad production was 8.8%, the median dyad production was 0.6%, and no dyads at all were observed in 13 plants. Triad production ranged from 0 to 6.9%, with a median of 0.8%, and 7 plants did not produce triads. Monad production ranged from 0 to 8.7% with a median of 0.9%, and 16 plants did not produce monads. Only one individual produced hexads (0.5%), as well as pentads (0.3%), and an additional two individuals also produced pentads at low frequencies ( %). Chromosome counts, FISH, and fertility in the second selfing generation Individual SP_042 had a high seed set (294), moderate pollen viability (62%), and DNA content 97% of a full hexaploid (2n = AABBCC). A set of 25 second-generation self-pollinated progeny from SP_042 were characterized by chromosome counts. The average chromosome number was 50, ranging from 46 to 54 with a standard deviation of 2.6. Three plants with.50 chromosomes were selected for investigation using FISH and are hereafter referred to as SP_042_01, SP_042_02, and SP_042_03 (Figure 6). SP42_01 had 53 chromosomes, compared with the expected chromosome complement for a allohexaploid plant of 54 (2n = AABBCC = 54). SP42_01 was missing four 278 A. S. Mason et al.

7 Figure 4 Distribution of pollen viability in 109 plants derived from self-pollination of a hybrid (A1) resulting from the cross (B. napus 3 B. carinata) 3 B. juncea, with overlaid density distribution (solid line) and normal distribution curve (dotted line). A-genome chromosomes and one C-genome chromosome and had an additional four recombinant chromosomes (Table 1 and Figure 6). This plant produced three seeds from 27 self-pollinated flowers (Table 1). SP42_02 had an estimated 54 chromosomes, but was missing 11 A-genome chromosomes and gained an additional six C-genome chromosomes and five recombinant chromosomes (Table 1). This plant set two seeds from 18 self-pollinated flowers (Table 1). SP42-03 had an estimated 53 chromosomes, was missing seven A-genome chromosomes and three B-genome chromosomes, and gained an additional four C-genome chromosomes and five recombinant chromosomes (Table 1). This plant was the most fertile of the three plants with a seed set of 124 seeds from 73 selfpollinated flowers (1.7 seeds/flower). Representative chromosome spreads for each of SP42-01, SP42-02, and SP42-03 are shown in Figure 6 with chromosomes from each genome labeled by FISH and recombinant chromosomes labeled. Correlations between fertility measures and DNA content/allele inheritance No significant association was observed between the loss of chromosome segments and seed set or pollen viability in the SP population (one-way ANOVA). A weak but significant correlation (one-way ANOVA, P = 0.02, r 2 = 0.12) existed between DNA content as estimated by flow cytometry and seed set, excluding the plant with abnormally low DNA content (Figure 2). No significant association between DNA content and pollen viability was observed, although pollen viability and seed set in the subset of the SP population assessed for DNA content were significantly associated (one-way ANOVA, P = 0.01, r 2 = 0.17). Discussion In this study, we assessed fertility, chromosome transmission, and genomic stability in self-pollinated progeny of a near-allohexaploid Brassica hybrid, using a combination of high-throughput molecular marker genotyping, flow cytometry, fertility measurements, and cytogenetics techniques. Approximately one-third of the self-pollinated progeny retained all parental A- and C-genome alleles, and unbiased allelic segregation was observed for.90% of the loci assessed. Surprisingly, this is higher than observed for previously generated B. napus mapping populations. For example, 22 49% of alleles showed segregation distortion in three well-characterized doubled-haploid-derived populations of B. napus (J. Wang et al. 2011). This difference in segregation bias may be related to the use of microspore culture to produce these mapping populations, rather than fertilization and embryo development in planta. Genomic regions conferring better microspore survival in culture may be selected for during the novel pressures of the culture process, resulting in allele segregation distortion. DNA content followed a normal distribution about the A1 plant mean, and several plants were highly self-fertile, setting hundreds of self-pollinated seeds. SNP data indicated that there was relatively little homeologous recombination across most of the genome, and allele transmission followed patterns consistent with normal Mendelian segregation (Figure 3). This represents an improvement over previously detected meiotic stability in synthetic Brassica types (Song et al. 1995; Szadkowski et al. 2010; Tian et al. 2010; Xiong et al. 2011; Zou et al. 2011). This effect may be related to the heterozygosity of the initial hexaploid hybrid used in this study (2n = A j A n B j B c C n C c ), which could facilitate rapid Allohexaploid Brassica Genetics 279

8 Figure 5 Distribution of self-pollinated seed set in 109 plants derived from self-pollination of a hybrid (A1) resulting from the cross (B. napus 3 B. carinata) 3 B. juncea, with overlaid density distribution (solid line). allelic selection for meiotic stability and fertility in resulting progeny. Despite the confounding presence of the A n 7-C n 6 translocation and putative A10 duplication in the B. napus parental line used to generate this population, suggestive initial results for DNA content, allele inheritance, and fertility were obtained. Taken together, these results suggested that several self-progeny of the interspecific hybrid had normal or near-normal, substantially un-rearranged chromosome complements. This is a promising result supporting the development of stable allohexaploid Brassica lines using this approach. In our study, there was no relationship between missing chromosome fragments and fertility in the SP population, and only a weak predictive correlation (11%) between DNA content and self-seed set. Although Nicolas et al. (2012) observed significantly reduced fertility in progeny of Brassica allohaploids (AC) with increasing loss of chromosome segments, the higher genome redundancy of the allohexaploids (AABBCC) may have protected against this effect. Loss of A7 and C6 chromosome regions in the allohexaploid population was also not significantly different from expected proportions for progeny resulting from selfing of a translocation heterozygote (12.5%), suggesting that rearrangement events were not discriminated against; similar results were observed by Nicolas et al. (2009) in allohaploid B. napus. Surprisingly, there was no selection pressure for or against loss of univalent chromosomes A j 2 and A j 6 in the allohexaploid progeny, and univalent chromosome A j 9 was present more often than predicted by chance (Figure 3). This contrasts with previous observations in Brassica of reduced (,50%) transmission of monosomic C-genome chromosomes in AA + C addition lines (Heneen et al. 2012) and biased transmission of univalents in AAC hybrid types (Leflon et al. 2006) and may be related to the higher ploidy level or genome structure of the allohexaploid hybrid types. Selection pressure to retain alleles conferring beneficial meiotic stability may also be operating to retain univalent chromosomes or chromosome regions. A number of QTL with effects on meiotic behavior have been identified in B. napus (Liu et al. 2006), including the major locus PrBn (Jenczewski et al. 2003) with quantitative effects on crossover frequency between nonhomologous chromosomes (Nicolas et al. 2009). Recently, presence of chromosomes C6 and C9 and retention of additional univalent chromosomes were also found to increase crossover frequency between homologous A-genome chromosomes in AA + C hybrids (Suay et al. 2014). Little is known about nonhomologous chromosome pairing control in B. juncea and B. carinata, although some evidence for a genetic locus in B. juncea with a major effect on nonhomologous pairing has been obtained (Prakash 1974). Preferential retention of A j 9 (and of univalents A j 2andA j 6relativetoprevious studies) in our population may be related to the presence of alleles for meiotic stability or fertility on those chromosomes or to the effect of additional univalent chromosomes on genome-wide meiotic behavior. Higher chromosome numbers in the three second-generation progeny did not correspond to euploid AABBCC chromosome complements (Table 1 and Figure 6). In fact, these progeny showed evidence of large-scale recombination between all three genomes (Table 1 and Figure 6), as well as loss of A-genome chromosomes but putative retention of additional sets of putative C-genome homeologs (Figure 6). The retention of four homeologous chromosomes for each set of homeologs but with loss or gain of homologous pairs (e.g., A1/A1/C1/C1 to A1/A1/A1/A1) was also reported by Szadkowski et al. (2010) and Xiong et al. (2011) in resynthesized 280 A. S. Mason et al.

9 Figure 6 FISH using genome labels for chromosomes in second-generation progeny derived from a near-allohexaploid Brassica (2n = AABBCC) plant [(B. napus 3 B. carinata) 3 B. juncea]: A-genome chromosomes are blue (DAPI, background stain), B-genome chromosomes are labeled red, and C-genome chromosomes are labeled green. Recombinant chromosomes are labeled with an asterisk, and the two parent genomes involved are noted. Images were taken at 350 magnification. Bar, 5 mm. B. napus. Similarly, Nicolas et al. (2009) observed no discrimination against rearranged chromosome complements transmitted from haploid B. napus, supporting lack of selection pressure for euploid, unrearranged chromosome complements in Brassica hybrids. Therefore, even though several plants in the SP generation hypothetically could have had euploid 2n = AABBCC chromosome complements (Figure 3), these plants may not have had a fertility advantage over their aneuploid counterparts: SP_042 was one of the most fertile of the SP generation, but gave rise to progeny with a standard deviation of 62.6 chromosomes. The SP42 plant is also predicted to have contained at least some chromosomal rearrangements and karyotype disturbances based on the commonality of the putatively rare A-B recombinant chromosomes observed in all three progeny using FISH (Figure 6), as well as the common loss of two to four pairs of A-genome chromosomes. A similar lack of correlation between fertility and aneuploidy, and even between aneuploidy and regularity of meiotic behavior, was also reported for Brassica hexaploids from the cross B. carinata 3 B. rapa (Iwasa 1964). The genome redundancy provided by the presence of three genomes may have reduced selection pressure for more normal chromosome complements, allowing retention of high frequencies of chromosome rearrangements, duplications, and deletions in the progeny. Generally, regular gamete formation was observed at the sporad stage. Only three plants (5%) produced any pentads or hexads, and these were at very low frequency ( %), suggesting that few laggard chromosomes were being lost due to formation of micronuclei during cytokinesis. This finding contrasts with Iwasa (1964), who found more than four nuclei in 5 18% of sporads in three 2n = AABBCC plants derived from the cross B. carinata 3 B. rapa, although no dyads, monads, or triads were observed in these allohexapoid hybrid types. On average, frequencies of unreduced gametes produced by the SP population were higher than in the parent Brassica species, but lower than in the B. napus 3 B. carinata hybrid types (Mason et al. 2011). Genetic factors related to unreduced gamete production in the parent plants (Mason et al. 2011) may have been partially transmissible to the majority of (B. napus 3 B. carinata) 3 B. juncea SP progeny, based on comparison with the results of Iwasa (1964). Several mutations resulting in high frequencies of unreduced gamete formation have been identified in Arabidopsis (D Erfurth et al. 2008; De Storme and Geelen 2011), and genetic Table 1 Number of chromosomes inherited per genome (A, B, or C) as detected by FISH in three second-generation plants derived from a near-allohexaploid 2n = AABBCC plant Plant A B C A/B A/C B/C Total Fertility Expected for an AABBCC allohexaploid SP42_ seeds/flower SP42_ seeds/flower SP42_ seeds/flower Allohexaploid Brassica Genetics 281

10 factors have been implicated in other species (Tavoletti et al. 1991). However, it is probable that univalent-induced meiotic restitution [the spread of univalents across the metaphase plate resulting in a restitution nucleus (Catcheside 1934; De Storme and Geelen 2013)] was responsible for the unreduced gamete that contributed to the allohexaploid plant A1. The absence of similarly high frequencies of unreduced gamete production in the SP population may support a mode of bivalent formation for most plants, as suggested by the molecular karyotyping results (Figure 3). We found that results from FISH in combination with molecular karyotyping were highly effective for determining detailed chromosome complements in cytogenetically complex hybrids, as has also been demonstrated in oats (Jellen et al. 1994), banana, and sugarcane (D Hont 2005). This analysis highlighted the dangers of using simple chromosome counts to assess meiotic stability of Brassica allohexaploids, as has been done in other studies (e.g., Howard 1942; Tian et al. 2010). In the future, molecular karyotyping with molecular cytogenetics may allow identification of genotypic and species-specific variability for meiotic stability in higher-ploidy Brassica. With molecular karyotyping, we could identify chromosomes and genomic regions with regular or irregular A- and C-genome allele transmission, and FISH helped to identify recombinant chromosomes and chromosome copy number. The addition of high-throughput molecular markers for the Brassica B genome would also be beneficial in characterization of Brassica allohexaploids to aid in the development of this new potential Brassica crop species. Production and characterization of additional hexaploid lines from different genotypes of B. napus, B. carinata, and B. juncea and intercrossing between different sources of allohexaploids (Geng et al. 2013) may also help in elucidating the complex interplay of genetic and genomic factors contributing to Brassica meiotic stability and contribute to the formation of stable, fertile allohexaploids. Acknowledgments We thank Rowan Bunch for Illumina HiScan SNP chip scanning. This work was supported by an Australian Research Council Discovery Early Career Researcher Award (DE ) and by Australian Research Council Linkage Project grants LP , DP , and LP Travel of A.S.M. to France to carry out molecular cytogenetic characterization was supported by an Australian Academy of Science France Australia Science Innovation Collaboration Early Career Fellowship. G.M.A. was supported by a scholarship from the Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brazil. Literature Cited Arumugam, N., A. Mukhopadhyay, V. Gupta, D. Pental, and A. K. Pradhan, 1996 Synthesis of hexaploid (AABBCC) somatic hybrids: a bridging material for transfer of tour cytoplasmic male sterility to different Brassica species. Theor. Appl. Genet. 92: Catcheside, D. G., 1934 The chromosomal relationships in the swede and turnip groups of Brassica. Ann. Bot. (Lond.) 48: Chen, S., M. N. Nelson, A. M. Chèvre, E. Jenczewski, Z. Li et al., 2011 Trigenomic bridges for Brassica improvement. Crit. Rev. Plant Sci. 30: Cousin, A., K. Heel, W. A. Cowling, and M. N. Nelson, 2009 An efficient high-throughput flow cytometric method for estimating DNA ploidy level in plants. Cytometry A 75: D Erfurth, I., S. Jolivet, N. Froger, O. Catrice, M. Novatchkova et al., 2008 Mutations in AtPS1 (Arabidopsis thaliana parallel spindle 1) lead to the production of diploid pollen grains. PLoS Genet. 4: e De Storme, N., and D. Geelen, 2011 The Arabidopsis mutant jason produces unreduced first division restitution male gametes through a parallel/fused spindle mechanism in meiosis II. Plant Physiol. 155: De Storme, N., and D. Geelen, 2013 Sexual polyploidization in plants: cytological mechanisms and molecular regulation. New Phytol. 198: D Hont, A., 2005 Unraveling the genome structure of polyploids using FISH and GISH: examples of sugarcane and banana. Cytogenet. Genome Res. 109: Dixon, G. R., 2007 Vegetable Brassicas and related crucifers, pp. 1:33 in Crop Production Science in Horticulture Series, edited by J. Atherton, and H. Rees. CAB International, Oxfordshire, UK. Fulton, T. M., J. Chunwongse, and S. D. Tanksley, 1995 Microprep protocol for extraction of DNA from tomato and other herbaceous plants. Plant Mol. Biol. Rep. 13: Geng, X. X., S. Chen, I. A. Astarini, G. J. Yan, E. Tian et al., 2013 Doubled haploids of novel trigenomic Brassica derived from various interspecific crosses. Plant Cell Tiss Org 113: Heneen, W. K., M. Geleta, K. Brismar, Z. Y. Xiong, J. C. Pires et al., 2012 Seed colour loci, homoeology and linkage groups of the C genome chromosomes revealed in Brassica rapa-b. oleracea monosomic alien addition lines. Ann. Bot. (Lond.) 109: Howard, H. W., 1942 The effect of polyploidy and hybridity on seed size in crosses between Brassica chinensis, B. carinata, amphidiploid B. chinensis-carinata and autotetraploid B. chinensis. J. Genet. 43: Howell, E. C., G. Barker, G. H. Jones, M. J. Kearsey, G. King et al., 2002 Integration of the cytogenetic and genetic linkage maps of Brassica oleracea. Genetics 161: Iwasa, S., 1964 Cytogenetic studies on the artificially raised trigenomic hexaploid hybrid forms in the genus Brassica. J. Faculty of Agriculture, Kyushu University 13: Jellen, E. N., B. S. Gill, and T. S. Cox, 1994 Genomic in-situ hybridization differentiates between A/D-genome and C-genome chromatin and detects intergenomic translocations in polyploid oat species (genus Avena). Genome 37: Jenczewski,E.,F.Eber,A.Grimaud,S.Huet,M.O.Lucaset al., 2003 PrBn, a major gene controlling homeologous pairing in oilseed rape (Brassica napus) haploids. Genetics 164: Leflon, M., F. Eber, J. C. Letanneur, L. Chelysheva, O. Coriton et al., 2006 Pairing and recombination at meiosis of Brassica rapa (AA) 3 Brassica napus (AACC) hybrids. Theor. Appl. Genet. 113: Leitch, A. R., and I. J. Leitch, 2008 Genomic plasticity and the diversity of polyploid plants. Science 320: Li, M., W. Qian, J. Meng, and Z. Li, 2004 Construction of novel Brassica napus genotypes through chromosomal substitution and elimination using interploid species hybridization. Chromosome Res. 12: A. S. Mason et al.

11 Liu, Z., K. Adamczyk, M. Manzanares-Dauleux, F. Eber, M.-O. Lucas et al., 2006 Mapping PrBn and other quantitative trait loci responsible for the control of homeologous chromosome pairing in oilseed rape (Brassica napus L.) haploids. Genetics 174: Mason, A. S., V. Huteau, F. Eber, O. Coriton, G. Yan et al., 2010 Genome structure affects the rate of autosyndesis and allosyndesis in AABC, BBAC and CCAB Brassica interspecific hybrids. Chromosome Res. 18: Mason, A. S., M. N. Nelson, G. J. Yan, and W. A. Cowling, 2011 Production of viable male unreduced gametes in Brassica interspecific hybrids is genotype specific and stimulated by cold temperatures. BMC Plant Biol. 11: 103. Mason, A. S., G. Yan, W. A. Cowling, and M. N. Nelson, 2012 A new method for producing allohexaploid Brassica through unreduced gametes. Euphytica 186: Mason, A. S., J. Batley, P. E. Bayer, A. Hayward, W. A. Cowling et al., 2014 High-resolution molecular karyotyping uncovers pairing between ancestrally-related Brassica chromosomes. New Phytol. DOI: /nph Meng, J., S. Shi, L. Gan, Z. Li, and X. Qu, 1998 The production of yellow-seeded Brassica napus (AACC) through crossing interspecific hybrids of B. campestris (AA) and B. carinata (BBCC) with B. napus. Euphytica 103: Morinaga, T., 1934 Interspecific hybridisation in Brassica VI. The cytology of F 1 hybrids of B. juncea and B. nigra. Cytologia (Tokyo) 6: Nelson, M. N., A. S. Mason, M.-C. Castello, L. Thomson, G. Yan et al., 2009 Microspore culture preferentially selects unreduced (2n) gametes from an interspecific hybrid of Brassica napus L. 3 Brassica carinata Braun. Theor. Appl. Genet. 119: Nicolas, S. D., M. Leflon, H. Monod, F. Eber, O. Coriton et al., 2009 Genetic regulation of meiotic cross-overs between related genomes in Brassica napus haploids and hybrids. Plant Cell 21: Nicolas, S. D., H. Monod, F. Eber, A. M. Chevre, and E. Jenczewski, 2012 Non-random distribution of extensive chromosome rearrangements in Brassica napus depends on genome organization. Plant J. 70: Osborn, T. C., D. V. Butrulle, A. G. Sharpe, K. J. Pickering, I. P. Parkin et al., 2003 Detection and effects of a homeologous reciprocal transposition in Brassica napus. Genetics 165: Otto, S. P., and J. Whitton, 2000 Polyploid incidence and evolution. Annu. Rev. Genet. 34: Pradhan, A., J. A. Plummer, M. N. Nelson, W. A. Cowling, and G. Yan, 2010 Successful induction of trigenomic hexaploid Brassica from a triploid hybrid of B. napus L. and B. nigra (L.) Koch. Euphytica 176: Prakash, S., 1974 Probable basis of diploidization of Brassica juncea Coss. Can. J. Genet. Cytol. 16: Rahman, M. H., 2001 Production of yellow-seeded Brassica napus through interspecific crosses. Plant Breed. 120: Roberts, A. V., 2007 The use of bead beating to prepare suspensions of nuclei for flow cytometry from fresh leaves, herbarium leaves, petals and pollen. Cytometry A 71: Sjödin, C., and K. Glimelius, 1989 Brassica naponigra, a somatic hybrid resistant to Phoma lingam. Theor. Appl. Genet. 77: Song, K. M., P. Lu, K. L. Tang, and T. C. Osborn, 1995 Rapid genome change in synthetic polyploids of Brassica and its implications for polyploid evolution. Proc. Natl. Acad. Sci. USA 92: Suay, L., D. Zhang, F. Eber, H. Jouy, M. Lodé et al., 2014 Crossover rate between homologous chromosomes and interference are regulated by the addition of specific unpaired chromosomes in Brassica. New Phytol. 201: Szadkowski, E., F. Eber, V. Huteau, M. Lodé, C. Huneau et al., 2010 The first meiosis of resynthesized Brassica napus, a genome blender. New Phytol. 186: Tavoletti, S., A. Mariani, and F. Veronesi, 1991 Phenotypic recurrent selection for 2n pollen and 2n egg production in diploid alfalfa. Euphytica 57: Tian, E., Y. Jiang, L. Chen, J. Zou, F. Liu et al., 2010 Synthesis of a Brassica trigenomic allohexaploid (B. carinata 3 B. rapa) de novo and its stability in subsequent generations. Theor. Appl. Genet. 121: U. N., 1935 Genome-analysis in Brassica with special reference to the experimental formation of B. napus and peculiar mode of fertilization. Jap. J. Bot. 7: Wang, J., D. J. Lydiate, I. P. Parkin, C. Falentin, R. Delourme et al., 2011 Integration of linkage maps for the amphidiploid Brassica napus and comparative mapping with Arabidopsis and Brassica rapa. BMC Genomics 12: 101. Wang, X. W., H. Z. Wang, J. Wang, R. F. Sun, J. Wu et al., 2011 The genome of the mesopolyploid crop species Brassica rapa. Nat. Genet. 43: Xiong, Z. Y., R. T. Gaeta, and J. C. Pires, 2011 Homoeologous shuffling and chromosome compensation maintain genome balance in resynthesized allopolyploid Brassica napus. Proc. Natl. Acad. Sci. USA 108: Zonneveld, B. J. M., I. J. Leitch, and M. D. Bennett, 2005 First nuclear DNA amounts in more than 300 angiosperms. Ann. Bot. (Lond.) 96: Zou, J., J. Zhu, S. Huang, E. Tian, Y. Xiao et al., 2010 Broadening the avenue of intersubgenomic heterosis in oilseed Brassica. Theor. Appl. Genet. 120: Zou, J., D. H. Fu, H. H. Gong, W. Qian, W. Xia et al., 2011 De novo genetic variation associated with retrotransposon activation, genomic rearrangements and trait variation in a recombinant inbred line population of Brassica napus derived from interspecific hybridization with Brassica rapa. Plant J. 68: Communicating editor: A. Houben Allohexaploid Brassica Genetics 283